Along the same vein is the problem of false positives in header searches because the headers contains the text content of the message in a header added by a Spam detection product like SpamAssassin used by a domain host.  Below is an example:

X-Ham-Report: Spam detection software,
 running on the system "host215.hostmonster.com",
 has NOT identified this incoming email as spam.  The original message has
 been attached to this so you can view it or label similar future email.  If
 you have any questions,
 see root\@localhost for details.  Content preview:  AUCTION ENDS TODAY!!
 TIME IS TICKING!! GET YOUR BIDS IN!! DON'T MISS IT!! HOTTEST DEALS GOING ON
 NOW!!! GOING ONCE GOING TWICE SOLD!! ... 

When I search one or more folders for specific word(s) or phrase(s), I find that I occasionally get "false positives"--messages that do not contain the search string in the message body. In nearly all cases these "false positives" are messages that have attachments. I've also noticed that this seems more prone to occur when the search string is fairly short--say fewer than six characters. Based on the results of a limited number of experiments, I've confirmed that the search is finding random occurrences of the search string in the part of the message containing the encoded binary attachment and considers these successful "finds", even though the search string is not contained in the body of the message.

Since it seems like it would be only very rarely that someone would need to find a specific text string in an encoded attachment, my suggestion is to revise PMail to: 1)not allow any searches of encoded attachments or 2)give the user the option to prohibit or allow such searches.

Thanks for your consideration,

Jim Wolfe